For submission to 2016 International Conference on Sustainable Development (ICSD)

AN INTEGRATIVE DOUBLE SKIN FAÇADE DAMPER SYSTEM FOR STRUCTURAL

SAFETY AND ENERGY EFFICIENCY

Rui ZHANG1) and Tat S. FU2)

1) PhD Candidate, Department of Civil and Environmental Engineering,
University of New Hampshire, Durham, New Hampshire, USA, rjp74@wildcats.unh.edu
2) Assistant Professor, Department of Civil and Environmental Engineering,
University of New Hampshire, Durham, New Hampshire, USA, tat.fu@unh.edu

Rui Zhang, PhD candidate, University of New Hampshire (corresponding author)

rjp74@wildcats.unh.edu
(603) 609 - 8669
Department of Civil and Environmental Engineering,
W183 Kingsbury Hall, 33 Academic Way,
Durham, NH 03824

Tat S. Fu, Assistant Professor, University of New Hampshire

1
ABSTRACT:
A double skin façade (DSF) system consists of two “skin” layers of glasses with a cavity
between the skins. The skins and cavity help insulate and ventilate buildings (air is allowed to
flow in the cavity). This project integrates mass dampers and double skin façades by enabling
movements in the outer skin. The DSF dampers can dampen structural vibrations during
earthquakes and strong winds without adding extra weight (i.e., separate mass dampers) to the
structure. Additionally, by being able to adjust the cavity size between the two skins, ventilation
rate can be controlled to improve energy efficiency. The synergy of the proposed system can
lead to buildings that are structurally safe, energy efficient, and ultimately sustainable. In this
study, the energy impact of the integrated system was first investigated. The airflow speed
inside the cavity can be controlled by adjusting the cavity size. Controlling airflow leads to
different control strategies to increase or decrease the solar heat gain in heating or cooling
seasons, respectively. The integrated system was also tested for mitigating structural response
under earthquakes. In the experimental study, a façade damper system with optimized damper
parameters was installed on a six-story shear structure and shake table tests were performed
on the DSF damper, a conventional tuned mass damper (TMD) and an uncontrolled system.
The experiment result showed that various DSF damper configurations significantly reduced
structural vibration and inter-story drift compared to the uncontrolled system and outperformed
the TMD system. This research utilized DSF which is an established architecture competent,
using an integrated approach to balance the needs of building hazard safety and energy saving
and achieved a cost-effective and sustainable design solution.

2
INTRODUCTION
An integrative system is recently proposed1 by authors by combining double skin façades and
mass dampers (Figure 1). A double skin façade (DSF) system consists of two “skin” (i.e.,
external) layers of glasses with a cavity between the skins. Compared to a single glass skin
system, this system is more insulating and its cavity allows airflow to help ventilate a building.
Mass dampers can damp and reduce structural vibrations during earthquakes and strong winds
by moving in a way that reduce the primary structure’s motions. Using façade systems as mass
dampers can eliminate the need to add extra mass (i.e., separate mass dampers) to the
structure. In terms of energy efficiency, the added mobility of façades (via actuators, see Figure
2) can lead to innovative ways to adjust airflow and environmental conditions in the building.
The synergy of the integrated system comes not only from its individual functions but also its
utilization. Structural control systems are used only during the infrequent recurrence of strong
motions, while environmental control systems are in constant use to provide continuous comfort
for building occupants. This integrated system would perform environmental control most of the
time and switch to structural control when needed, thus providing a synergistic dual-purpose
system to improve building energy efficiency and enhance structural and life safety.

Figure 1. Double skin façade (DSF)

damper system; the outer skin Figure 2. Motions of the movable double
moves as a mass damper. skin façade system.

The DSF-damper system proposes an interesting and challenging structural system. Given that
the DSF covers all exterior surfaces of a building, multiple dampers and actuators are required
to connect the façade system throughout the building, as shown in Figure 1. The motions of
these dampers are also coupled because they are connected to the same DSF system. This

1
Fu, Tat S., and Rui Zhang. "Integrating Double-Skin Façades and Mass Dampers for Structural Safety
and Energy Efficiency." Journal of Architectural Engineering (2016): 04016014.

3
coupled multiple damper system is more difficult for engineers to design because of the
coupling effect and the various locations of the individual dampers, but it can be less disruptive
for architectural design because there are no additional bulky damper masses in the building.
Because the façade system is heavy, the proposed DSF-damper system can be more massive
than a traditional mass-damper system. DSFs are traditionally stationary, but the movements for
mass dampers mobilize the façade system and offers novel methods to reduce building energy
consumptions. As shown in Figure 3, the outer façade is attached to the primary building with a
rail system consist of bearings, dampers and actuators. By moving the outer skin, the cavity
between the two skins can be adjusted according to internal building conditions (heating/cooling
needs) and external conditions (outside temperatures, wind conditions, etc.). In a typical
building, HVAC (heating, ventilation, air conditioning) systems typically draw the largest amount
of energy compared to other building functions. Given that cavity size influences cavity airflow
and the insulating effectiveness of the façade system, the movable DSFs affords an efficient
way to regulate temperatures and ventilation in the building. During hot summer and cold winter
days, a highly insulating façade system is preferred to shield building occupants from the
discomforting temperatures outside. On the other hand, in a cool summer night, a building that
is still fairly warm from
daytime heat gain can be
cooled faster with a less
insulating façade.
While mass dampers and
DSFs are not new
technologies individually,
marrying the two is
innovative and provides a
synergy between the fields of
Civil Engineering and
Architecture. This paper
focuses on validating using
DSFs as structural control
devices. Specifically, both
passive damping and active
control strategies are
investigated for mitigation of Figure 3. Details of the connections between the primary
building and the movable façade: the wide flange rail system
structural vibration.
is attached with bearings, dampers and actuators to allow and
control linear motions.

Background
Mass Dampers

A mass damper is a secondary mass attached to the primary structure that is designed to affect
(reduce) structural motions. One of the most common mass damper system is the tuned mass
damper (TMD). A TMD can only suppress the response of a primary system in narrow
frequency band and, therefore, is ineffective for excitation in other frequency ranges. Thus, a
multiple tuned mass damper (MTMD) system was introduced to increase the robustness by

4
tuning MTMDs to wider frequency bands. Igusa and Xu2 and Yamaguchi and Harnpornchai3 first
proposed a MTMD to compensate for the sensitivity of a single TMD to the uncertain natural
frequencies of the building system. The MTMD was later extensively studied by Abe and
Fujino4 and Kareem and Kline5. Most of these other 1990s studies6,7,8 concentrate the multiple
dampers in one floor or use just a single degree-of-freedom (DOF) model of the primary system;
some other research9,10 has had dampers on several floors of a multistory building. In contrast,
the proposed DSF damper system herein has dampers in all floors because the DSF covers the
building surfaces and is structurally supported throughout the building. Previous work by the
authors11 includes a distributed mass damper system that also places mass dampers in all
floors of a building. However, the DSF damper system is different than the author’s previous
work because the DSF dampers connected multiple floors, coupling the motions of several
floors. Figure 4 illustrates the difference between the single TMD, a single-floor MTMD, the
distributed mass damper system and the proposed DSF damper systems.

Figure 4. Mass damper systems

2
Igusa, T., and K. Xu. "Vibration control using multiple tuned mass dampers." Journal of sound and
vibration 175, no. 4 (1994): 491-503.
3
Yamaguchi, Hiroki, and Napat Harnpornchai. "Fundamental characteristics of multiple tuned mass
dampers for suppressing harmonically forced oscillations." Earthquake engineering & structural dynamics
22, no. 1 (1993): 51-62.
4
Abe, Masato, and Yozo Fujino. "Dynamic characterization of multiple tuned mass dampers and some
design formulas." Earthquake engineering & structural dynamics 23, no. 8 (1994): 813-835.
5
Kareem, Ahsan, Tracy Kijewski, and Yukio Tamura. "Mitigation of motions of tall buildings with specific
examples of recent applications." Wind and structures 2, no. 3 (1999): 201-251.
6
Li, Chunxiang. "Performance of multiple tuned mass dampers for attenuating undesirable oscillations of
structures under the ground acceleration."Earthquake Engineering & Structural Dynamics 29, no. 9
(2000): 1405-1421.
7
Li, Chunxiang. "Optimum multiple tuned mass dampers for structures under the ground acceleration
based on DDMF and ADMF." Earthquake engineering & structural dynamics 31, no. 4 (2002): 897-919.
8
Zuo, Lei, and Samir A. Nayfeh. "Optimization of the individual stiffness and damping parameters in
multiple-tuned-mass-damper systems." Journal of Vibration and Acoustics 127, no. 1 (2005): 77-83.
9
Bergman, L. A., D. M. McFarland, J. K. Hall, E. A. Johnson, and A. Kareem. "Optimal distribution of
tuned mass dampers in wind-sensitive structures." In Structural Safety and Reliability, pp. 95-102. ASCE,
1989.
10
Chen, Genda, and Jingning Wu. "Optimal placement of multiple tune mass dampers for seismic
structures." Journal of Structural Engineering 127, no. 9 (2001): 1054-1062.
11
Fu, Tat S., and Erik A. Johnson. "Distributed mass damper system for integrating structural and
environmental controls in buildings." Journal of Engineering Mechanics 137, no. 3 (2010): 205-213.

5
Moon12 first proposed the idea using a double skin façade (DSF) as a mass damper. In that
study, the author showed vibration reduction by modeling the primary structure as a single DOF
system and the DSF damper as the 2nd DOF element. By modeling the overall system as a
2DOF system, Moon’s model is the most basic TMD system. There are some weaknesses to
that model. First, it is not applicable to typical buildings with DSFs where the primary structure is
generally a multiple DOF system. Second, because the DSF covers all exterior surfaces of a
building, multiple dampers are required to connect the façade system throughout the building as
shown in Figure 4. This paper investigates these issues by formulating a multiple DOF structural
model with multiple dampers connecting the DSF. Another similar system is Obayashi
Corporation’s “Hula” mass damper13. In the Hula concept system, the outer walls of the building
are attached with dampers. This system is similar to a configuration of the DSF system in this
study: when the entire DSF is considered rigid and a single mass damper. However, given that
a rigid single DSF damper (spanning the entire height of a building) is unpractical for tall
buildings, configurations of multiple DSF dampers are studied in this paper.
Double Skin Façade

In common DSF systems, a DSF is an envelope sealed with two glass skins with ventilation
openings near the bottom and top of the façade system. As illustrated in Figure 5, when solar
radiation strikes the external glass skin, it largely passes straight through the external glazing
because glasses have a high transmittance to shortwave (solar) radiation. A portion of the solar
radiation is absorbed by the glass skin and causes the glass to be heated. After crossing the
external skin, the solar radiation heats the air in the cavity between the two glass skins and
then passes through the second glass skin to enter the building. Again, a portion of the solar
radiation is absorbed by the interior skin. As the cavity gets warm, they (like any warm object)
emits thermal radiation in long wavelengths. Given that glasses have low transmittance to
longwave radiation, the heat is trapped between the two glass skins and thus amplifies the heat
gain in the cavity. The process of glass trapping solar heat gain is called the greenhouse effect.

Figure 5. Heat Transfer in DSF.

12
Moon, Kyoung Sun. "Tall building motion control using double skin façades." Journal of architectural
engineering 15, no. 3 (2009): 84-90.
13
Obayashi Corporation (2012). “Hula Mass Damper System.” Accessed in March 11, 2013
<http://www.obayashi.co.jp/service_and _technology/related/tech60> (in Japanese).

6
In DSF systems, there is another natural effect that influences the heating and cooling aspects:
the stack effect. In a tall space, hot air rises to the top of the space because it is lighter than cold
air. At the top of the space, the concentrated hot air creates a higher air pressure compared to
the outside air and this higher pressure pushes air outward. The outflows of air on the top of the
space forces cooler, denser air enter from the bottom. The rate of airflow, Q, induced by the
stack effect can be computed using:

Ti - TO (1)
Q = CA 2 gh
Ti
where A is the flow area; C is a discharge coefficient; g is the gravitational acceleration; h is the
height of the space; Ti is the average inside temperature and To is the average outside air
temperature. According to (1), the airflow rate will increase in a DSF cavity when the cavity
height grows and/or the temperature difference between outside air and air inside the cavity
increases.
There are different control strategies for DSF systems in cooling and heating seasons. In
cooling seasons, outside temperatures are generally higher than indoor temperatures, and
DSFs are rapidly heated up by solar radiation and the greenhouse effect. In terms of thermal
comfort and cooling energy saving, it would be advantageous to ventilate DSF cavities with a
large airflow rate to prevent cavity air being overheated. Figure 6(a) showed that increases in air
change rates would lower cooling loads. Figure 6(a) also shows that decreases in air change
rates would lower heating loads. During heating seasons, the external temperature is cooler
than the indoor temperature; and a thermal buffer zone (such as a DSF with a higher cavity
temperature) can help insulate the building from the cold outside environments. Low air change
rates would cause cavity temperature to raise14.
(a) Energy cost VS airflow rate (b) Airflow rate VS cavity depth
100

7
80
5350
Annual Cooling Load (KWh/m )

Annual Heating Load (KWh/m )

2

Air Change per Hour (ACH)

50
6
60
Air Velocity (m/s)

5
40
5300

4
20

45 Air Velocity with 8% Opening Area

Annual Heating Load 3 ACH with 8% Opening Area
5250 0
Annual Cooling Load ACH with 16% Opening Area
ACH with 24% Opening Area
2
1 1.2 1.4 1.6 1.8 2 100 200 400 600 800 1000
Air Change per Hour (ACH) 4 Cavity Depth (mm)
x 10
Figure 6. (a) Energy cost compared to airflow; (b) airflow compared to DSF cavity depth (source
15 16
data from Torres et al. and Tascón ).

14
Gratia, Elisabeth, and André De Herde. "Greenhouse effect in double-skin facade." Energy and
buildings 39, no. 2 (2007): 199-211.
15
Torres, Mauricio, Pere Alavedra, Amado Guzmán, Eva Cuerva, Carla Planas, Raquel Clemente, and
Vanessa Escalona. "Double skin facades-Cavity and Exterior openings Dimensions for Saving energy on
Mediterranean climate." IBPSA Proceedings of Building Simulation (2007): 198-20.
16
Tascón, Mauricio Hernandez. Experimental and computational evaluation of thermal performance and
overheating in double skin facades. University of Nottingham, 2008.

7
Depending on the DSF configurations (e.g., ventilation opening sizes, façade height), there are
different depths for each DSF system to achieve fastest or slowest air flow rates. Figure 6(b)
shows different airflow rate curves for various cavity depths in different DSF configurations15,16.
Large depths are preferred for slow air flow rates that could cause cavity to heat up quickly and
decrease heating loads in buildings. To obtain the fastest airflow rates and reduce cooling
loads, an optimal depth is desired (e.g., 400-600 mm in Figure 6(b)). Tascón16 showed that
airflow was restricted in cavities smaller than the optimal depth while a cavity larger than the
optimal depth would form low speed turbulence eddies that decrease the overall airflow. Figure
6 shows that cooling and heating seasons demand different cavity depths to reduce energy
consumption, highlighting the potential of the proposed movable DSF system in which DSF
cavity depths can be adjusted.

Structural Model and Formulation

The equations of motions for the n-story structure with TMD or DSF dampers can be expressed
as Equation (2),

M!x! + Cx! + Kx = -M1!x!g (2)

In this equation, M is the mass matrix, C is the damping matrix and K is the stiffness matrix, x =
[x1 x2 … xn x1d x2d … xnd ]T, xi and xid are the ith floor displacement and the ith damper displacement
relative to the ground, respectively. !x!g is the ground acceleration and 1 is a column vector of
ones. The difference between the traditional TMD system and the DSF system are illustrated in
Table 1.

15
Torres, Mauricio, Pere Alavedra, Amado Guzmán, Eva Cuerva, Carla Planas, Raquel Clemente, and
Vanessa Escalona. "Double skin facades-Cavity and Exterior openings Dimensions for Saving energy on
Mediterranean climate." IBPSA Proceedings of Building Simulation (2007): 198-20.
16
Tascón, Mauricio Hernandez. Experimental and computational evaluation of thermal performance and
overheating in double skin facades. University of Nottingham, 2008.

8
 TMD DSF

é[M s ]n´n 0ù é[M s ]n´n 0 n´l ù

M=ê M=ê
ë 0 m d úû ë 0l´n [M d ]l´l úû

é[K ] - kdù é [K11 ]n´n [K12 ]n´l ù

K = ê s nd´n
ë -k
ú
kd û
K=ê
[
ë (K12 ) l´n
T
] [K 22 ] l´l úû
Table 1. The equation of motion for TMD system (left column) and DSF damper system (right
column).

For both systems, damping matrix C takes a similar form as K. Here, mi and m id are the
masses of the ith floor and m id of the ith damper, respectively; ki and k id are the stiffness
coefficients of the ith floor and between the ith floor and the corresponding damper, respectively.
For TMD systems, the damper is connected to the top story of the building, the mass matrix and
stiffness matrices of the primary structure are as following:

ém1 0 ! 0ù ék1 + k 2 - k2 0 ! 0 ù
ê0 ê -k ú
m2 0 úú ê 2 k 2 + k3 - k3 0 ú
Ms = ê , Ks = ê 0 # " ú.
ê" # " ú ê ú
ê ú ê " - k n-1 k n-1 + k n - kn ú
ë0 0 ! mn û
ê 0 ! 0 - kn k n + k d úû
ë

9
For DSF damper systems, the DSF dampers connected multiple floors, coupling the motions of
several floors. Consider a multiple DSF damper system with n stories and l dampers where n = r
l. Here, r is a scalar vale, implying that l is chosen to be able to divide n. For example, a six
story structure can have three DSF dampers with each damper attached to two floors. This
system with evenly divided DSF dampers is selected for simplicity and scalability in simulations.

ém1d 0 ! 0ù
ê d ú
0 m 0ú
Md = ê 2
,
ê " # " ú
ê ú
ëê 0 0 ! mld ûú

ék1 + k 2 + k1d - k2 0 ! 0 ù
ê ú
ê - k2 k 2 + k 3 + k 2d - k3 0 ú
K 11 = ê 0 # " ú
ê ú
ê " - k n -1 k n -1 + k n + k nd-1 - kn ú
ê 0 ! 0 - kn k n + k nd úû ,
ë

T
ék1d k 2d ! k rd 0 0 ! 0 ! 0 0 0 0 ù
ê ú
0 0 ! 0 k rd+1 k rd+ 2 ! k 2dr 0 0 0 0
K 12 = -ê ú
ê" # " ú
ê dú
êë 0 0 ! 0 0 0 ! 0 ! k (dl -1) r +1 k (dl -1) r +1 ! k lr úû
,

and

é r d ù
êå k i 0 0 ! 0 ú
ê i =1 2r ú
ê 0 0 ú
ê åk
i = r +1
i
d
0
ú
K 22 =ê 3r ú.
ê 0
ê
0 åk
i = 2 r +1
i
d
0 ú
ú
ê " # " ú
ê lr ú
ê 0
êë
0 0 ! å k id ú
i = ( l -1) r +1 úû

The corresponding state space representation of (2) is

éx! ù é 0 I ù éx ù é 0 ù
ê!x!ú = ê- M -1K - M -1Cú êx! ú + ê- 1ú !x!g (3)
ë û ë ûë û ë û

10
or z = Az + Bu , where u is the system input vector. The output vector y = Cyz + Dyu can include,
for example, floor drifts, velocities and/or absolute accelerations depending on Cy and Dy. For a
passively controlled system, u depends only on the ground acceleration. A Kanai-Tajimi
stochastic model17 of earthquake ground motion is used as the excitation: a low-pass filtered
Gaussian white noise with filter

2z gw g s + w g2
F(s) = (4)
s 2 + 2z gw g s + w g2

where wg = 17 rad/s and zg = 0.3 to mimic the frequency content of four historical ground
motions (1940 El Centro, 1968 Hachinohe, 1995 Kobe and 1994 Northridge; Ramallo et al.18).
This stochastic model with white noise was the excitation input was used to optimize the DSF
units’ parameters.

Experimental setup
In order to check the performance of the proposed DSF mass damper system, a six-story shear
scaled model were fabricated to serve as a base structure. Three configurations of the DSF
mass damper system and a TMD system were installed on the base structure and tested
respectively. This section discusses the experimental structure and the damper mechanisms.
Structure setup
The experimental structure was a six-story shear structure as shown in Figure 7. It had a
dimension of 0.46 m (18 inches) in length, 0.46 m (18 inches) in width and 1.85 m (73 inches) in
height and was mounted on a shake table at the University of New Hampshire. The structure
was a metal frame structure consists of four steel columns with rectangle cross sections and six
aluminum plates equally spaced to represent floors. Aluminum plates were held by friction using
four extra steel stocks at corners to keep the plate perpendicular to the column. Each story had
identical height of 30.5 cm (12 inches) and identical weight of 23.8 kg (52.5 lb). The
fundamental period was designed to be 0.6 seconds to represent this six-story building
structure.

17
Soong, Tsu T., and Mircea Grigoriu. "Random vibration of mechanical and structural systems." NASA
STI/Recon Technical Report A 93 (1993).
18
Ramallo, J. C., E. A. Johnson, and B. F. Spencer Jr. "“Smart” base isolation systems." Journal of
Engineering Mechanics 128, no. 10 (2002): 1088-1099.

11
 Figure 7. The six-story experiment structure

Three DSF damper configurations are tested: one-, two- and three-damper configurations as
shown on Table 2 and Figure 8. In the one-damper configurations, the façade spanned and
connected to all six floors and, thus, moved as an entire piece. In the two- and three-damper
configurations, each of the DSF dampers spanned three or two floors, respectively.

Number of dampers
Floor
1 2 3

1 1st
1st damper
2
damper
3 1st 2nd
damper damper
4
2nd
5 3rd
damper
6 damper

Table 2. DSF mass damper configurations

12
 Figure 8. DSF damper configurations: three dampers configuration (left), two dampers
configuration (middle) and one damper configuration (right).

In this experiment, the DSF damper had a 10% damper mass ratio (relative to the primary
structural floor mass). Typically, mass damper systems use damper mass ratios less than 2%.
This is because that damper masses are kept small to avoid introducing a large amount of extra
weight to the structure. In the proposed system, DSFs are part of the existing architectural
systems in a building and, thus, are not additional weights to the structure. Building façade
systems and their weights can vary significantly from one building to another. Given that they
form the entire surfaces of buildings, building façades are heavy, especially in DSF in which
there are two layers of glasses and a dedicated support structural system for the outer glass
layer. In this study, the DSF dampers account for 10% of the overall structural mass.
Each DSF damper consisted of a two fiber glass panels and served as the movable outer skin
of DSF. A set of extension springs and a sliding dual shaft linear guide as shown in Figure 9.
The fiber glass panels on each side of the building were connected by a horizontal T-bar which
attached to the linear guide’s movable carriage. The linear guides were oriented along the
shake table stroke direction. A pair of extension springs were connected between the T-bar and
a set of aluminum frames. The vertical component of the aluminum frame was made adjustable
thus different sizes of springs could fit in. In the three-damper configuration shown in Figure 8,
façades spanned two floors could move entirely as one unit. The weight of DSF damper
consisted of the two fiber glass panels, a moving carriage on the linear shaft and a T-bar. The
DSF mass damper weighed 2.22 kg (4.91 lb) and was approximately 10% of damper mass ratio
(relative to the primary structural mass). The frequency of the damper could be adjusted by
using different springs.

13
 Figure 9. Details of the DSF mass damper mechanism

Tuned mass damper (TMD)

In this study, a single TMD was attached to the top story of the experiment building (see Figure
10). The TMD system has the same mechanical component compared to the DSF system
except that the façade panels were replaced with equivalent weights (steel stocks) lumped on
the linear guide on the structure’s top story. The damper mass ratio (10%) of the TMD system
was kept the same as the DSF damper system. The frequency of the damper could be adjusted
by using different springs. Compared to the DSF system setup, the structure with the TMD had
all the static parts of dampers remained on 1st -5th floor, while all the moveable damper mass
was removed at those floors. The carriage was locked on the linear guide to prohibit any motion.

Figure 10. Details of the tuned mass damper (TMD)

14
Comparisons were made between three cases: 1) an uncontrolled structure enclosed with a
static DSF, 2) a structure enclosed with a static DSF and equipped with a TMD on the top floor,
3) a structure with the proposed movable DSF damper system. However, since the façade
panels were designed to span multiple floors under the experiment setup, the façade panels
had to be removed to avoid coupling multiple floors. To make a fair comparison to the DSF
system, equivalent weights of the façade panels were installed on each floor for the TMD and
uncontrolled structure as static weights.

Test apparatus
The testing apparatus was a seismic shake table manufactured by MTS Systems Corporation.
This shake table (Figure 11) consists of a 2.5m × 2.5m aluminum sliding plate and a 97.8 kN
(22-kip) linear hydraulic actuator with a 15.2cm (6 inch) stroke. The shake table can be
displacement-controlled through an LVDT (linear variable differential transformer) or
acceleration-controlled by using an accelerometer. Accelerations were recorded during the
testing. Six Microstrain G-Link-LXRS wireless accelerometers were installed on each floor of the
building. The wireless accelerometers had a 12-bit resolution, ±2 g range and the sampling rate
was set to 256 Hz. A WSDA-1000 data aggregator was used to collect the acceleration test
data. In addition, floor displacements were measured via digital image correlation (DIC)
technique. A target painted with speckle pattern was mounted on each floor (see Figure 2.9),
Two Photron high speed cameras were placed on the side of structure to track the motion of the
target at a sampling frequency of 60Hz. The captured frames were processed through DIC
software to obtain the floor displacements. Inter-story drifts were then derived.

Figure 11. Shake table.

15
Result Discussion
Five experimental test cases (an uncontrolled structure, an optimized TMD system, and three
configurations of the DSF damper systems) were performed under five historic earthquake
records (i.e., 1992 Erzincan, 1994 Northridge, 1995 Kobe, 1999 Jiji and 2011 New Hall). The
absolute accelerations at all six floors were measured. Figure 12 shows a comparison of the top
floor (6th floor) time history acceleration of the uncontrolled, TMD and three-damper DSF
systems under the Northridge and Jiji earthquakes. The earthquake excitation length was about
30 seconds for the Northridge earthquake and 50 seconds for the Jiji earthquake. Cleary shown
in Figure 12, both the TMD and DSF systems reduced significant amounts of vibrations
compared to the uncontrolled structure. The structure with the TMD reduced the top-floor
vibration by 79.80% under Northridge and by 56.60% under Jiji. Meanwhile, the DSF dampers
outperformed the TMD by reducing responses by 83.37% under Northridge and by 64.56%
under Jiji.

Figure 12. Top floor acceleration response under Northridge (left) and Jiji Earthquake (Right).

Tables 3 and 4 show the reduction in both inter-story drifts and acceleration, respectively. Three
different DSF damper configurations and a conventional TMD were compared to the
uncontrolled structure performed under the five scaled historic earthquake excitations.

Average RMS inter-story drift of all six floors (mm)

Excitation Uncontrolled TMD DSF three-damper DSF two-damper DSF one-damper
Drift Drift Reduct. Drift Reduct. Drift Reduct. Drift Reduct.
Northridge 1.38 1.16 -16.18% 1.19 -14.15% 1.12 -18.93% 1.14 -17.29%
Kobe 1.98 1.68 -15.36% 1.71 -31.85% 1.21 -51.97% 1.17 -53.29%
New Hall 0.85 0.65 -23.12% 0.52 -58.04% 0.49 -61.10% 0.48 -61.71%
Jiji 2.06 1.76 -14.53% 1.74 -15.64% 1.76 -14.78% 1.65 -20.07%
Erzincan 3.20 1.96 -38.78% 1.12 -65.03% 1.24 -61.22% 1.24 -61.30%
Average 1.89 1.44 -21.59% 1.26 -36.94% 1.16 -41.60% 1.14 -42.73%
Table 3. Average RMS Inter-story drift of the DSF damper systems compare to TMD system and
uncontrolled structure (RMS: root mean square).

16
 Average RMS acceleration of all six floors (g)
Excitation Uncontrolled TMD DSF three-damper DSF two-damper DSF one-damper
Accel. Accel. Reduct. Accel. Reduct. Accel. Reduct. Accel. Reduct.
Northridge 0.0203 0.015 -26.11% 0.0145 -28.57% 0.013 -35.96% 0.0126 -37.93%
Kobe 0.0875 0.0695 -20.57% 0.0347 -60.34% 0.0371 -57.60% 0.0352 -59.77%
New Hall 0.1146 0.0998 -12.91% 0.0832 -27.40% 0.0754 -34.21% 0.0745 -34.99%
Jiji 0.0406 0.0241 -40.64% 0.0181 -55.42% 0.0202 -50.25% 0.016 -60.59%
Erzincan 0.1349 0.1141 -15.42% 0.0272 -79.84% 0.0333 -75.32% 0.0336 -75.09%
Average 0.0796 0.0645 -23.13% 0.0355 -50.31% 0.0358 -50.67% 0.0344 -53.67%
Table 4. Average RMS acceleration of the DSF damper systems compare to TMD system and
uncontrolled structure (RMS: root mean square).

As shown in Table 3, all DSF damper systems reduced significant amounts of inter-story drifts
compared to the uncontrolled structure without any dampers, and the overall performance of
DSF outperforms that of the conventional TMD system. Among the DSF configurations, the one-
damper configuration was overall the best performing system; and the two-damper configuration
was a close second. The one-damper case reduced an averaging 42.73% of inter-story drift and
53.67% of floor acceleration compared to the uncontrolled structure. In terms of the average
inter-story drift reduction, the one- and two-damper configurations outperformed the TMD
system in all the excitations, while three-damper case outperformed the TMD case in all
excitations except Northridge. In terms of the average acceleration reduction, all DSF damper
systems outperformed the TMD system in all five earthquake excitations.

Conclusion

In this paper, an innovative system integrating double skin façades (DSF) and mass damper
was analyzed. By motorizing DSFs, the movable façade system can improve energy efficiency
by adjusting the airflow between the skin façades. Previous studies showed that heating loads
typically decrease with large DSF cavity depths (slow airflow in the cavity) and cooling loads are
minimized with small DSF cavity depths (fast airflow in the cavity). In addition to energy
efficiency, the resulting DSF mass damper system can significantly reduce structural motions
when subject to earthquake excitation. Experiments were conducted in which a scaled structure
with DSF dampers on a shake table was subjected to historical earthquake records. The DSF
mass damper system was shown to significantly reduce structural motions. Additionally, the
proposed system was shown to mostly outperform a conventional tuned mass damper system
with an equivalent damper mass ratio.

Acknowledgment

The authors gratefully acknowledge the partial support by the National Science Foundation
under grant CMMI 12-28074. Any opinions, findings, and conclusion or recommendations
expressed herein are those of the authors and do not necessarily reflect the views of the
National Science Foundation.

17